From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Wed, 19 Jul 2000 14:51:06 +0200
From: Lucio De Re <lucio@proxima.alt.za>
To: 9fans@cse.psu.edu
Subject: [9fans] Re: =?iso-8859-1?Q?=5B9fans=5D_Re:_R=E9f=2E_:_Re:_=5B9fans=5D_mothra?=
Message-ID: <20000719145105.G3081@cackle.proxima.alt.za>
References: <41256921.0048318B.00@SNPAR12.> <009901bff17d$37d44f10$62356887@HWTPC>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <009901bff17d$37d44f10$62356887@HWTPC>; from Howard Trickey on Wed, Jul 19, 2000 at 08:31:09AM -0400
Topicbox-Message-UUID: e4d35e0e-eac8-11e9-9e20-41e7f4b1d025

On Wed, Jul 19, 2000 at 08:31:09AM -0400, Howard Trickey wrote:
> 
> Yes..  Now how do you draw a line from
> the end of the troff pipe back into tbl in
> your diagram?  You need the full power
> of the layout engine available to the table
> processer, as sizing depends on the layout
> of nested tables in a very insidious way.

That seems unavoidable.  And you've been there, I haven't :-)  But it
strikes me as a recursive descent problem that should really have only
one solution, or, at worst, scroll bars :-)

> and the real fun begins when a cell is allocated
> a certain width, and the contents don't fit.
> Or when the various width and pad specs
> of the whole nested set of tables are
> impossible to satisfy simultaneously.

Perhaps some scaling knobs in the right places...

> Then you have to "give up" in exactly the
> same way as Netscape and IE...
> (I seem to have the phrase burned on
> my brain, don't I?)

... rather than rely on previous foolishness :-)

> And then try to handle images of unknown
> size arriving later, and try to minimize the
> number of unnecessary trial layouts.
> 
Except for the need to match prior art (which I understand and hope to
be right in disregarding) I believe we can solve all these problems,
if for no other reason that typography is not a recent invention.  If
we aim at elegance rather than "accuracy" (intentional scare quotes),
and we make the product readily available, we may just swing enough
decision makers to make a difference.

If lynx wasn't so hideously ugly, I believe it would have made its
mark in this sense too.

> Most of my layout bugs were due to
> tables.
> 
How many were actual errors, versus mere divergence from a crooked de
facto standard?

> - the fool
> 
Your experience is undoubtably invaluable.

++L